U.S. patent application number 11/236719 was filed with the patent office on 2006-04-20 for method for determining a failure of a manufacturing condition, system for determining a failure of a manufacuring condition and method for manufacturing an industrial product.
Invention is credited to Tomomi Ino, Yukihiro Ushiku.
Application Number | 20060085165 11/236719 |
Document ID | / |
Family ID | 36177568 |
Filed Date | 2006-04-20 |
United States Patent
Application |
20060085165 |
Kind Code |
A1 |
Ushiku; Yukihiro ; et
al. |
April 20, 2006 |
Method for determining a failure of a manufacturing condition,
system for determining a failure of a manufacuring condition and
method for manufacturing an industrial product
Abstract
A method for determining a failure of a manufacturing condition,
includes creating waveforms implemented by respective first data
strings of first characteristic variables corresponding to
operation parameter data of manufacturing apparatuses which execute
manufacturing processes of products under respective manufacturing
conditions for the products; classifying the first data strings
that are analogous to each other into groups based on a correlation
of the waveforms; creating a first visualized data table
visualizing magnitude correlations between the first characteristic
variables for each of the groups; adding second data strings of
second characteristic variables to the groups, the second
characteristic variables corresponding to workmanship data
representing measurement and inspection results of the products;
and creating a second visualized data table visualizing magnitude
correlations between the second characteristic variables for each
of the groups.
Inventors: |
Ushiku; Yukihiro;
(Yokohama-shi, JP) ; Ino; Tomomi; (Yokohama-shi,
JP) |
Correspondence
Address: |
FINNEGAN, HENDERSON, FARABOW, GARRETT & DUNNER;LLP
901 NEW YORK AVENUE, NW
WASHINGTON
DC
20001-4413
US
|
Family ID: |
36177568 |
Appl. No.: |
11/236719 |
Filed: |
September 28, 2005 |
Current U.S.
Class: |
702/183 ;
714/E11.207 |
Current CPC
Class: |
G05B 23/024
20130101 |
Class at
Publication: |
702/183 |
International
Class: |
G06F 11/30 20060101
G06F011/30 |
Foreign Application Data
Date |
Code |
Application Number |
Sep 30, 2004 |
JP |
P2004-287948 |
Claims
1. A computer implemented method for determining a failure of a
manufacturing condition, comprising: creating a plurality of
waveforms implemented by respective first data strings of first
characteristic variables corresponding to operation parameter data
of a plurality of manufacturing apparatuses which execute a
plurality of manufacturing processes of a plurality of products
under respective manufacturing conditions for the products;
classifying the first data strings that are analogous to each other
into a plurality of groups based on a correlation of the waveforms;
creating a first visualized data table for each of the groups, the
first visualized data table visualizing magnitude correlations
between the first characteristic variables; adding second data
strings of second characteristic variables to the groups, the
second characteristic variables corresponding to workmanship data
representing measurement and inspection results of the products;
and creating a second visualized data table for each of the groups,
the second visualized data table visualizing magnitude correlations
between the second characteristic variables.
2. The method of claim 1, further comprising: extracting a target
group from the groups based on the magnitude correlations between a
target second characteristic variable in the second characteristic
variables by using the second visualized data table; and extracting
a target first characteristic variable from the first
characteristic variables of the target group based on the magnitude
correlations between the first characteristic variables by using
the first visualized data table.
3. The method of claim 1, wherein the first characteristic
variables are deviations from average values of the operation
parameter data.
4. The method of claim 1, wherein the groups are classified based
on a correlation coefficient for all combinations of the
products.
5. The method of claim 1, wherein the first and second graphic
patterns represent the magnitude correlations by one of sizes of
circles, lengths of bars, and densities of gray scales.
6. The method of claim 1, wherein the operation parameter data
include at least one of pressure, gas flow rate, wafer temperature,
and input radio frequency power in at least one process chamber of
the manufacturing apparatuses.
7. The method of claim 1, wherein the products are semiconductor
devices.
8. The method of claim 7, wherein the workmanship data include
quality control characteristics after completion of each of the
manufacturing processes and electrical characteristics of the
semiconductor devices.
9. A system for determining a failure of a manufacturing condition,
comprising: a plurality of monitor units configured to acquire
operation parameter data of a plurality of manufacturing
apparatuses which execute a plurality of manufacturing processes of
a plurality of products under respective manufacturing conditions
for the products; an inspection tool configured to acquire
workmanship data representing measurement and inspection results of
the products; a waveform creation module configured to create a
plurality of waveforms implemented by first data strings of first
characteristic variables corresponding to the operation parameter
data for each of the products; a classification module configured
to classify the first data strings that are analogous to each other
into a plurality of groups based on a correlation of the waveforms;
a table creation module configured to create a first visualized
data table for each of the groups and a second visualized data
table by adding second data strings of second characteristic
variables corresponding to the workmanship data to the groups, the
first visualized data table visualizing magnitude correlations
between the first characteristic variables, the second visualized
data table visualizing magnitude correlations between the second
characteristic variables; and an internal memory configured to
store the operation parameter data, the workmanship data, the first
and second data strings, the groups, and the first and second
visualized data tables.
10. The system of claim 9, wherein the first characteristic
variables are deviations from average values of the operation
parameter data.
11. The system of claim 9, wherein the groups are classified based
on a correlation coefficient for all combinations of the
products.
12. The system of claim 9, wherein the first and second graphic
patterns represent the magnitude correlations by one of sizes of
circles, lengths of bars, and densities of gray scales.
13. The system of claim 9, wherein the operation parameter data
include at least one of pressure, gas flow rate, wafer temperature,
and input radio frequency power in at least one process chamber of
the manufacturing apparatuses.
14. The system of claim 9, wherein the products are semiconductor
devices.
15. The system of claim 14, wherein the workmanship data include
quality control characteristics after completion of each of the
manufacturing processes and electrical characteristics of the
semiconductor devices.
16. A method for manufacturing an industrial product, comprising:
executing manufacturing processes of a plurality of products under
respective manufacturing conditions for the products; acquiring
operation parameter data of a plurality of manufacturing
apparatuses which execute the manufacturing processes, the
operation parameter data corresponding to the manufacturing
conditions; creating a plurality of waveforms implemented by first
data strings of first characteristic variables corresponding to the
operation parameter data for each of the products; classifying the
first data strings that are analogous to each other into a
plurality of groups based on a correlation of the waveforms;
creating a first visualized data table for each of the groups, the
first visualized data table visualizing magnitude correlations
between the first characteristic variables; acquiring workmanship
data representing measurement and inspection results of the
products; adding second data strings of second characteristic
variables to the groups, the second characteristic variables
corresponding to the workmanship data; creating a second visualized
data table for each of the groups, the second visualized data table
visualizing magnitude correlations between the second
characteristic variables; extracting a target group from the groups
based on the magnitude correlations between a target second
characteristic variable in the second characteristic variables by
using the second visualized data table; extracting a target first
characteristic variable from the first characteristic variables of
the target group based on the magnitude correlations between the
first characteristic variables by using the first visualized data
table; and executing a target manufacturing process by determining
measures for a target manufacturing condition corresponding to the
target first characteristic variable at the next production.
17. The manufacturing method of claim 16, wherein the first
characteristic variables are deviations from average values of the
operation parameter data.
18. The method of claim 16, wherein the groups are classified based
on a correlation coefficient for all combinations of the
products.
19. The method of claim 16, wherein the first and second graphic
patterns represent the magnitude correlations by one of sizes of
circles, lengths of bars, and densities of gray scales.
20. The method of claim 16, wherein the operation parameter data
include at least one of pressure, gas flow rate, wafer temperature,
and input radio frequency power in at least one process chamber of
the manufacturing apparatuses.
Description
CROSS REFERENCE TO RELATED APPLICATIONS AND INCORPORATION BY
REFERENCE
[0001] This application is based upon and claims the benefit of
priority from prior Japanese Patent Application P2004-287948 filed
on Sep. 30, 2004; the entire contents of which are incorporated by
reference herein.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The present invention relates to a method and a system for
determining a failure of a manufacturing condition of a
semiconductor device.
[0004] 2. Description of the Related Art
[0005] Along with advances in multifunction semiconductor devices
such as a semiconductor integrated circuit, reduction of pattern
dimensions and large-scale integration are in constant demand. It
is necessary to manufacture such a semiconductor device in a
plurality of chip regions on a semiconductor substrate with a
uniform performance and a high manufacturing yield.
[0006] In a manufacturing method for the semiconductor device,
various manufacturing processes are used. In order to improve the
manufacturing yield of the semiconductor device, it is necessary to
improve a yield rate for each of the manufacturing processes.
Therefore, failure determination of manufacturing conditions which
affect the yield rate for each of the manufacturing processes is
important.
[0007] To determine a failure of a manufacturing condition, a
correlation analysis of characteristic variable data which
represents various operation parameters of manufacturing
apparatuses is implemented. Usually, products of the semiconductor
device having similar behaviors of characteristic variable data are
classified into the same group. The characteristic variable data in
each group are superimposed with each other so that the correlation
for the characteristic variable data is visually determined. With
such a method, characteristic operation parameters of the
manufacturing apparatuses in each group can be understood. However,
it is difficult to understand a correlation between workmanship of
the manufacturing processes evaluated for each group and each
operation parameter. Hence, it is difficult to visually determine
operation parameters of manufacturing apparatuses which affect
product performance.
SUMMARY OF THE INVENTION
[0008] A first aspect of the present invention inheres in a
computer implemented method for determining a failure of a
manufacturing condition, including creating a plurality of
waveforms implemented by respective first data strings of first
characteristic variables corresponding to operation parameter data
of a plurality of manufacturing apparatuses which execute a
plurality of manufacturing processes of a plurality of products
under respective manufacturing conditions for the products;
classifying the first data strings that are analogous to each other
into a plurality of groups based on a correlation of the waveforms;
creating a first visualized data table for each of the groups, the
first visualized data table visualizing magnitude correlations
between the first characteristic variables; adding second data
strings of second characteristic variables to the groups, the
second characteristic variables corresponding to workmanship data
representing measurement and inspection results of the products;
and creating a second visualized data table for each of the groups,
the second visualized data table visualizing magnitude correlations
between the second characteristic variables.
[0009] A second aspect of the present invention inheres in system
for determining a failure of a manufacturing condition, including a
plurality of monitor units configured to acquire operation
parameter data of a plurality of manufacturing apparatuses which
execute a plurality of manufacturing processes of a plurality of
products under respective manufacturing conditions for the
products; an inspection tool configured to acquire workmanship data
representing measurement and inspection results of the products; a
waveform creation module configured to create a plurality of
waveforms implemented by first data strings of first characteristic
variables corresponding to the operation parameter data for each of
the products; a classification module configured to classify the
first data strings that are analogous to each other into a
plurality of groups based on a correlation of the waveforms; a
table creation module configured to create a first visualized data
table for each of the groups and a second visualized data table by
adding second data strings of second characteristic variables
corresponding to the workmanship data to the groups, the first
visualized data table visualizing magnitude correlations between
the first characteristic variables, the second visualized data
table visualizing magnitude correlations between the second
characteristic variables; and an internal memory configured to
store the operation parameter data, the workmanship data, the first
and second data strings, the groups, and the first and second
visualized data tables.
[0010] A third aspect of the present invention inheres in a method
for manufacturing an industrial product, including executing
manufacturing processes of a plurality of products under respective
manufacturing conditions; acquiring operation parameter data of a
plurality of manufacturing apparatuses which execute the
manufacturing processes, the operation parameter data corresponding
to the manufacturing conditions for the products; creating a
plurality of waveforms implemented by first data strings of first
characteristic variables corresponding to the operation parameter
data for each of the products; classifying the first data strings
that are analogous to each other into a plurality of groups based
on a correlation of the waveforms; creating a first visualized data
table for each of the groups, the first visualized data table
visualizing magnitude correlations between the first characteristic
variables; acquiring workmanship data representing measurement and
inspection results of the products; adding second data strings of
second characteristic variables to the groups, the second
characteristic variables corresponding to the workmanship data;
creating a second visualized data table for each of the groups, the
second visualized data table visualizing magnitude correlations
between the second characteristic variables; extracting a target
group from the groups based on the magnitude correlations between a
target second characteristic variable in the second characteristic
variables by using the second visualized data table; extracting a
target first characteristic variable from the first characteristic
variables of the target group based on the magnitude correlations
between the first characteristic variables by using the first
visualized data table; and executing a target manufacturing process
by determining measures for a target manufacturing condition
corresponding to the target first characteristic variable.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] FIG. 1 is a block diagram showing an example of a
configuration of a system for determining a failure of a
manufacturing condition according to an embodiment of the present
invention.
[0012] FIG. 2 is a diagram showing examples of first data strings
according to the embodiment of the present invention.
[0013] FIG. 3 is a diagram showing examples of first data string
waveforms according to the embodiment of the present invention.
[0014] FIG. 4 is a diagram showing an example of classification of
groups according to the embodiment of the present invention.
[0015] FIG. 5 is a diagram showing an example of a first visual
data table according to the embodiment of the present
invention.
[0016] FIG. 6 is a diagram showing examples of second data strings
according to the embodiment of the present invention.
[0017] FIG. 7 is a diagram showing an example of a second visual
data table according to the embodiment of the present
invention.
[0018] FIG. 8 is a flowchart showing an example of a method for
determining a failure of a manufacturing condition according to the
embodiment of the present invention.
[0019] FIG. 9 is a diagram showing another example of a first
visual data table according to the embodiment of the present
invention.
[0020] FIG. 10 is a diagram showing another example of a second
visual data table according to the embodiment of the present
invention.
[0021] FIG. 11 is a diagram showing an example of a first visual
data table according to other embodiment of the present
invention.
[0022] FIG. 12 is a diagram showing an example of a second visual
data table according to the other embodiment of the present
invention.
[0023] FIG. 13 is a diagram showing another example of a first
visual data table according to the other embodiment of the present
invention.
[0024] FIG. 14 is a diagram showing another example of a second
visual data table according to the other embodiment of the present
invention.
DETAILED DESCRIPTION OF THE INVENTION
[0025] Various embodiments of the present invention will be
described with reference to the accompanying drawings. It is to be
noted that the same or similar reference numerals are applied to
the same or similar parts and elements throughout the drawings, and
the description of the same or similar parts and elements will be
omitted or simplified.
[0026] As shown in FIG. 1, a system for determining a failure of a
manufacturing condition according to an embodiment of the present
invention includes a visualization processing unit 30, an input
unit 50, an output unit 52, an external memory 54, a manufacturing
control system 60, a manufacturing information database 62, a
plurality of manufacturing apparatuses 64a, 64b, 64n, a plurality
of monitor units 66a, 66b, . . . , 66n, an inspection tool 68, and
the like. Additionally, the visualization processing unit 30
includes an input module 32, a waveform creation module 34, a
classification module 36, a table creation module 38, an output
module 40, an internal memory 42, and the like.
[0027] Manufacturing processes for a plurality of products are
executed under respective manufacturing conditions of the
manufacturing apparatuses 64a, 64b, . . . , 64n. The visualization
processing unit 30 acquires operation parameter data corresponding
to the manufacturing conditions. Waveforms of first data strings of
first characteristic variables corresponding to the operation
parameter data for each of the plurality of products are created.
Based on a correlation between the waveforms, the first data
strings analogous to each other are classified into groups. A first
visualized data table for each of the groups is created. In the
first visualized data table, magnitude correlations between the
first characteristic variables are visualized by a first graphic
pattern. Moreover, workmanship data representing results of
measurements and inspections of the plurality of products are
acquired. Second data strings of second characteristic variables
corresponding to the workmanship data are added to the groups.
Thereafter, a second visualized data table for each of the groups
is created. In the second visualized data table, magnitude
correlations between the second characteristic variables are
visualized by a second graphic pattern.
[0028] The manufacturing control system 60 and the manufacturing
information database 62 are connected to the visualization
processing unit 30 through a local area network (LAN) 70. The
plurality of manufacturing apparatuses 64a, 64b, . . . , 64n, and
the plurality of monitor units 66a, 66b, . . . , 66n provided to
the plurality of manufacturing apparatuses 64a, 64b, . . . , 64n,
respectively, are connected to the manufacturing control system 60
through the LAN 70.
[0029] For example, manufacturing processes for a semiconductor
device, as a product, are executed by using the plurality of
manufacturing apparatuses 64a, 64b, . . . , 64n under corresponding
manufacturing conditions. Various sensors attached to each of the
plurality of manufacturing apparatuses 64a, 64b, . . . , 64n
monitor operation parameters, during processing, which correspond
to the manufacturing conditions. The manufacturing apparatuses 64a,
64b, . . . , 64n include a chemical vapor deposition (CVD)
apparatus, a vapor deposition apparatus, a dry etching apparatus,
an ion implantation apparatus, a photolithography system and the
like. In the case of a CVD apparatus, the sensors include a
pressure gauge, a flow meter, a thermometer, a radio frequency (RF)
power meter and the like.
[0030] The plurality of monitor units 66a, 66b, . . . , 66n
provided for the plurality of manufacturing apparatuses 64a, 64b, .
. . , 64n, respectively, acquire operation parameter data monitored
by the sensors and transmit the operation parameter data to the
manufacturing control system 60. The operation parameter data
includes, for example, pressure, gas flow rate, temperature, input
RF power and the like within a process chamber of a CVD
apparatus.
[0031] The manufacturing control system 60 collects the operation
parameter data transmitted from the plurality of monitor units 66a,
66b, . . . , 66n. The manufacturing control system 60 stores the
product numbers of the plurality of products processed by the
plurality of manufacturing apparatuses 64a, 64b, . . . , 64n, and
the operation parameter data corresponding to each of the products
in the manufacturing information database 62.
[0032] Additionally, the inspection tool 68 is connected to the
manufacturing control system 60 through the LAN 70. The inspection
tool 68 acquires workmanship data for each of the products by
measuring and inspecting quality control characteristics after
completion of each manufacturing process and electrical
characteristics after completion of the entire manufacturing
processes.
[0033] The quality control characteristics of an insulating film or
the like deposited by a CVD apparatus, for example, include film
thickness, in-plane uniformity, refraction index, the number of
particles, deposition rate, and the like. The electrical
characteristics of a field-effect transistor (FET), for example,
include a threshold voltage, an on-state current, a
transconductance, and the like. The manufacturing control system 60
stores the workmanship data, such as the quality control
characteristics and the electrical characteristics, measured and
inspected by the inspection tool 68, together with the product
numbers in the manufacturing information database 62.
[0034] From the manufacturing information database 62, the input
module 32 of the visualization processing unit 30 acquires the
operation parameter data of the plurality of manufacturing
apparatuses 64a, 64b, . . . , 64n which have executed the
manufacturing processes for the plurality of products. Further,
from the manufacturing information database 62, the input module 32
acquires the workmanship data which represents the measurements and
inspection results of the plurality of products.
[0035] The waveform creation module 34 creates a first data string
for each of the product numbers. The first data string has a string
of values of a plurality of first characteristic variables
corresponding to the operation parameter data of the plurality of
manufacturing apparatuses 64a, 64b, . . . , 64n. For example, first
characteristic variables PFa, PFb, PFc, PFd, and PFe corresponding
to the operation parameter data, such as pressure, gas flow rate,
temperature, input RF power and the like of a CVD apparatus, are
laid out in a line for each of the product numbers, as shown in
FIG. 2. Each of the first characteristic variables PFa, PFb, PFc,
PFd, and PFe is calculated, for example, as a deviation from an
average value of all of the plurality of products. Furthermore, as
shown in FIG. 3, the waveform creation module 34 creates waveforms
of the first data strings by depicting the first data string with
respect to the plurality of first characteristic variables for each
of the plurality of products as a line graph.
[0036] As shown in FIG. 4, the classification module 36 classifies
the waveforms of the first data strings that are be analogous to
each other based on a correlation between the waveforms of the
first data strings of the respective products, into a plurality of
groups Gr1 to Gr9. For example, using the created waveforms of the
first data strings, a correlation coefficient is calculated for all
combinations between the products. The waveforms of the first data
strings of the products which have a higher correlation coefficient
than a previously specified threshold value, are determined to be
analogous to one another. For example, specifying the threshold
value of the correlation coefficient to 0.8, the waveforms of the
first data strings are classified. Thus, the plurality of products
are classified into the plurality of groups Gr1 to Gr9 based on the
wavelengths of the first data strings.
[0037] As shown in FIG. 5, the table creation module 38 creates a
first visualized data table for indicating the plurality of first
characteristic variables PFa, PFb, PFc, PFd, and PFe using graphic
patterns, which visualizes magnitude correlations between the
plurality of groups Gr1 to Gr9. In the embodiment of the present
invention, circles are used as the graphic patterns. Each magnitude
of the plurality of first characteristic variables PFa, PFb, PFc,
PFd, and PFe is indicated by the size of the circle.
[0038] For example, with respect to each of the plurality of first
characteristic variables PFa, PFb, PFc, PFd, and PFe, a difference
between maximum and minimum values among the average values of
deviations in each of the plurality of groups Gr1 to Gr9 is divided
into four levels. Different sizes of the circles are assigned
respectively to the corresponding magnitudes of the average values
of the deviations.
[0039] The second data strings having strings of numerical values
of a plurality of second characteristic variables corresponding to
the workmanship data acquired by the input module 32 are added to
the first data strings of the products which are classified into
the plurality of groups Gr1 to Gr9, respectively. For example, as
shown in FIG. 6, the second characteristic variables RFa, RFb, RFc,
RFd, and RFe corresponding to the workmanship data, such as film
thickness, in-plane uniformity, refraction index, the number of
particles, deposition rate, and the like for the quality control
characteristics of CVD, are added to the first data strings. As the
second characteristic variables, the workmanship data, such as a
threshold voltage, an on-state current, a transconductance, and the
like for the electric characteristics of a FET can be used. As
shown in FIG. 7, the table creation module 38 creates the second
visualized data table where different sizes of the circles are used
as graphic patterns to indicate magnitudes of the plurality of
second characteristic variables RFa, RFb, RFc, RFd, and RFe in each
of the groups Gr1 to Gr9.
[0040] For example, with respect to each of the plurality of second
characteristic variables RFa, RFb, RFc, RFd, and Rfe, a difference
between maximum and minimum values among the average values of
deviations in each of the second characteristic variables RFa, RFb,
RFc, RFd, and Rfe is divided into four levels in each of the groups
Gr1 to Gr9. Different sizes of the circles are assigned
respectively to corresponding magnitudes of deviations.
[0041] The output module 40 transmits the first and second
visualized data tables to the output unit 52 so as to display the
first and second visualized data tables. In the first and second
visualized data tables, the magnitudes of the first and second
characteristic variables PFa, PFb, PFc, PFd, Pfe, and RFa, RFb,
RFc, RFd, Rfe are visually indicated by sizes of the graphic
patterns, respectively. Therefore, the workmanship data and the
operation parameter data of the manufacturing apparatuses can be
visually compared to each other in each of the plurality of groups
Gr1 to Gr9.
[0042] The internal memory 42 stores the operation parameter data
and the workmanship data acquired by the input module 32, the first
and second data strings and the groups created by the waveform
creation module 43 and classification module 36, respectively, the
first and second visualized data tables created by the table
creation module 38, and the like.
[0043] The input unit 50 refers to devices such as a keyboard and a
mouse. When an input operation is performed from the input unit 50,
corresponding key information is transmitted to the visualization
processing unit 30. The output unit 52 refers to a screen monitor,
such as a liquid crystal display (LCD), a light emitting diode
(LED) panel, an electroluminescent (EL) panel and the like. The
output unit 52 displays data tables processed by the visualization
processing unit 30 and graphic tables and the like acquired by the
same.
[0044] The external memory 54 stores a programs which allow the
visualization processing unit 30 to implement graphic processing,
classification by statistical operations, table creation, and the
like, for the acquired data. The internal memory 42 or the external
memory 54 of the visualization processing unit 30 temporarily
stores data obtained during a calculation and an analysis thereof
during the operation of the visualization processing unit 30.
[0045] As described above, the system for determining a failure of
a manufacturing condition according to the embodiment of the
present invention classifies the plurality of products into the
plurality of groups Gr1 to Gr9 based on the waveforms of the first
data strings created from the operation parameters of the
manufacturing apparatuses 64a, 64b, . . . , 64n which execute
manufacturing processes of the plurality of products. In each of
the plurality of groups Gr1 to Gr9, magnitude correlations between
the first and second characteristic variables corresponding to the
operation parameters of the manufacturing apparatuses and
workmanships of the products, respectively, are visualized by
graphic patterns.
[0046] Therefore, according to the embodiment of the present
invention, a relation between the workmanship of the products and
the operation parameters of manufacturing apparatuses 64a, 64b,
64n, which may not be understood heretofore, can be easily
recognized. Thus, it is possible to determine an operation
parameter of the manufacturing apparatuses 64a, 64b, . . . , 64n,
which cause a failure in the workmanship.
[0047] Next, a method for determining a failure of a manufacturing
condition according to the embodiment of the present invention is
described with reference to the flowchart shown in FIG. 8.
[0048] In step S100, manufacturing processes for the plurality of
products are executed using the plurality of manufacturing
apparatuses 64a, 64b, . . . , 64n shown in FIG. 1 under
corresponding manufacturing conditions. During the manufacturing
processes, operation parameters of the plurality of manufacturing
apparatuses 64a, 64b, . . . , 64n, which correspond to
manufacturing conditions, are monitored by the plurality of monitor
units 66a, 66b, . . . , 66n and stored in the manufacturing
information database 62, as the operation parameter data, through
the manufacturing control system 60.
[0049] In step S101, the operation parameter data is acquired by
the input module 32 of the visualization processing unit 30 from
the manufacturing information database 62.
[0050] In step S102, the waveform creation module 34 creates first
data string waveforms for each of the plurality of products from
the first data strings having a string of values of the plurality
of first characteristic variables PFa, PFb, PFc, PFd, and PFe
corresponding to the operation parameter data.
[0051] In step S103, the classification module 36 classifies the
first data strings that are analogous to one another into the
plurality of groups Gr1 to Gr9, based on correlations between the
first data string waveforms.
[0052] In step S104, for each of the plurality of groups Gr1 to
Gr9, the table creation module 38 creates a first visualized data
table which visualizes magnitude correlations of the plurality of
first characteristic variables PFa, PFb, PFc, PFd, and PFe using a
first graphic pattern.
[0053] The inspection tool 68 inspects and measures workmanship of
the plurality of products, such as quality control characteristics
after completion of each manufacturing process and electrical
characteristics after completion of the entire manufacturing
processes. The manufacturing control system 60 stores the
workmanship data measured by the inspection tool 68 into the
manufacturing information database 62.
[0054] In step S105, the input module 32 acquires a plurality of
second characteristic variables RFa, RFb, RFc, RFd, and RFe, which
correspond to the workmanship data, from the manufacturing
information database 62.
[0055] In step S106, second data strings having strings of values
of the plurality of second characteristic variables RFa, RFb, RFc,
RFd, and RFe are added to the plurality of groups Gr1 to Gr9,
respectively.
[0056] In step S107, table creation module 38 creates a second
visualized data table which indicates magnitude correlations of the
respective second characteristic variables RFa, RFb, RFc, RFd, and
RFe in each of the plurality of groups Gr1 to Gr9, by a second
graphic pattern.
[0057] In step. S108, based on the magnitude correlationships of a
target second characteristic variable in the plurality of second
characteristic variables RFa, RFb, RFc, RFd, and RFe, a target
group which is a determination target for a failure of a
manufacturing process, is extracted from the plurality of groups
Gr1 to Gr9. A target first characteristic variable which is the
cause of poor workmanship of the target group, is extracted from
the first characteristic variables PFa, PFb, PFc, PFd, and PFe of
the target group based on the magnitude correlations between the
first characteristic variables by using the first visualized data
table.
[0058] In step S109, based on a target operation parameter of the
manufacturing apparatus corresponding to the extracted target first
characteristic variable, measures to improve a manufacturing
condition of the manufacturing process are determined.
[0059] Thus, after revising the manufacturing conditions
corresponding to the operation parameters for the desired product
quality, manufacturing processes are execute at the next
production.
[0060] As a product example, during a deposition process of a
semiconductor device, the pressure, the flow rate of silane gas
(SiH.sub.4), the flow rate of hydrogen (H.sub.2), the wafer
temperature, the RF power, and the like, in a CVD chamber are
obtained as operation parameter data by a monitor unit of a CVD
apparatus. As shown in FIG. 9, a first visualized data table is
created for the groups Gr1 to Gr9 which are classified based on
waveforms of the first characteristic variable data strings
corresponding to the operation parameter data of the deposition
process. In the first visualized data table, magnitude correlations
between first characteristic variables are indicated by the size of
graphic patterns.
[0061] After the deposition process, a quality control process is
executed and the results (workmanship) of the deposition process
are measured. A film thickness, an in-plane uniformity, a
refraction index of a deposited film, a number of particles, a
deposition rate, and the like, are measured as the workmanship data
of the deposition process. Second characteristic variables
corresponding to the measured workmanship data are classified into
the groups Gr1 to Gr9 which correspond to the product numbers.
Thereafter, as shown in FIG. 10, the second visualized data table,
which indicates magnitude correlations between the second
characteristic variables by the sizes of the graphic patterns, is
created.
[0062] For example, when there is a problem in the deposition
process such that the number of particles is large and the yield
rate is decreased, the group, in which the large number of
particles are detected, is extracted from the second visualized
data table. As shown in FIG. 10, the graphic patterns indicate the
number of particles is large in the groups Gr2 and Gr3. Thus, the
groups Gr2 and Gr3 are visually recognized as the target groups for
determining a failure of a manufacturing condition.
[0063] Next, it is visually easily recognized that, in the first
visualized data table shown in FIG. 9, sizes of the graphic
patterns for flow rates of SiH.sub.4 and H.sub.2, and temperature,
of both the target groups Gr2 and Gr3 are smaller than other
groups. As a result, it is determined that the number of particles
increases under deposition conditions with low flow rates of
SiH.sub.4 and H.sub.2, and low wafer temperature. It can be
understood that increases in flow rates of SiH.sub.4 and H.sub.2 as
well as in wafer temperature are useful measures to reduce the
number of particles.
[0064] Furthermore, in order to obtain deposition conditions to
achieve a small film thickness, a good in-plane uniformity, and a
small number of particles, an group, in which sizes of graphic
patterns for a film thickness, an in-plane uniformity, and the
number of particles is small, are extracted from the second
visualized data table. As shown in FIG. 10, the group Gr7 is
visually recognized as the target group for determining a failure
of a manufacturing condition. It is determined from the first
visualized data table shown in FIG. 9 that the deposition process
of the target group Gr7 has a high flow rate of SiH.sub.4, a high
wafer temperature, and average values of other first characteristic
variables. Therefore, by adjusting set values of deposition
conditions of the CVD apparatus, a deposition process for a small
film thickness, a good in-plane uniformity and a small number of
particles is achieved.
[0065] In the method for determining a failure of a manufacturing
condition according to the embodiment of the present invention, a
relation between the workmanship data and the operation parameter
data of the manufacturing apparatuses is visually easily
recognized. Thus, it is possible to determine a failure of a
manufacturing condition which affects a yield rate of a
manufacturing process.
Other Embodiments
[0066] In the embodiment of the present invention, a manufacturing
process of a semiconductor device is described as an example.
However, it should be easily understood from the foregoing
descriptions that the present invention can also be applied to
manufacturing processes of industrial products such as automobiles,
chemicals, building components, liquid crystal devices, magnetic
recording mediums, optical recording mediums, thin film magnetic
heads, superconductor devices, and the like.
[0067] Additionally, in the embodiment of the present invention,
circles are used as graphic patterns in the first and second
visualized data tables. However, the graphic patterns are not
limited to circles. For example, as shown in FIGS. 11 and 12, bars
can be used as the graphic patterns in the first and second
visualized data tables. The magnitude correlations between
deviations of the first and the second characteristic variables are
expressed by the lengths of the bars. Alternatively, as shown in
FIGS. 13 and 14, a gray scale can be used as the graphic patterns
of the first and second visualized data tables. The magnitude
correlations between deviations of the first and second
characteristic variables are expressed by different densities of
the gray scale.
[0068] Various modifications will become possible for those skilled
in the art after storing the teachings of the present disclosure
without departing from the scope thereof.
* * * * *